Introduction
The Blue Swallow Hirundo atrocaerulea is restricted to sub-Saharan Africa, migrating between its breeding grounds in south-eastern and north-eastern South Africa, north-western Swaziland, eastern Zimbabwe, western Mozambique, southern and northern Malawi, north-eastern Zambia, southern Tanzania and the south-eastern Democratic Republic of the Congo (DRC) and non-breeding grounds in western Kenya, southern Uganda, north-western Tanzania and north-eastern DRC (Turner and Rose Reference Turner and Rose1989, Keith et al. Reference Keith, Urban and Fry1992). The birds occupy their breeding grounds from late-September to mid-April and their non-breeding grounds from late-April to late August or early-September (Keith et al. Reference Keith, Urban and Fry1992). The basic biology of the Blue Swallow is summarised in Keith et al. (Reference Keith, Urban and Fry1992) and Hockey et al. (Reference Hockey, Dean and Ryan2007).
The global Blue Swallow population, estimated at fewer than 1,500 pairs, is classified as ‘Vulnerable’ due to its small and rapidly declining population, resulting from the rapid reduction in the quantity and quality of its grassland and wetland habitat (BirdLife International 2000, 2008). The East African Blue Swallow population (Uganda, Kenya and Tanzania) is classified as ‘Endangered’ under East Africa regional Red Data criteria (Bennun and Njoroge Reference Bennun and Njoroge1996). The South African and Swaziland populations are both classified as ‘Critically Endangered’ (Evans and Barnes Reference Evans, Barnes and Barnes2000, Monadjem et al. Reference Monadjem, Boycott, Parker and Culverwell2003).
Current maps of the African distribution range of the Blue Swallow are only approximate (e.g. Turner and Rose Reference Turner and Rose1989). No quantitative assessment has previously been made of the decline in the Blue Swallow’s distribution range and population size. Similarly, no assessment has been made of what portion of the Blue Swallow population is located in formally, partially, or unprotected areas on both their breeding and non-breeding grounds. Also, no attempt has been made to determine the migratory routes between their breeding and non-breeding grounds. A better understanding of the current distribution range, migratory routes, population size and protection status would improve our ability to conserve the Blue Swallow and the grassland and wetland habitat on which it and other biodiversity depends.
Materials and methods
Distribution
All literature and other available information on the Blue Swallow have been collected over the 12 years of the present study. In addition, copies of all nest record cards of Blue Swallows were obtained from BirdLife South Africa, the Avian Demography Unit, and BirdLife Zimbabwe. The Endangered Wildlife Trust’s Blue Swallow Monitoring Database was also updated. Blue Swallow distribution data for South Africa and Zimbabwe, as contained in Allan and Earlé (Reference Allan, Earlé, A Harrison, Allan, Underhill, Herremans, Tree, Parker and Brown1997), was obtained from the Avian Demography Unit as ArcView 3.2 shapefiles. The Zimbabwe Department of National Parks and Wildlife Management provided all the information on Blue Swallows contained in their Field Cards. In 2005, a request to all birders to submit their sightings and any additional information they may have collected on Blue Swallows was posted on the Southern African Bird Net. All references used are provided in Appendix 1 in Supplementary Materials.
All of the above information was carefully scrutinised and the following data extracted per country: locality, number of nests recorded, number of skins collected, number of birds sighted, the reference, and any additional notes. Only original data were included. In keeping with Harrison et al. Reference Harrison, Allan, Underhill, Herremans, Tree, Parker and Brown1997, locality data were used to determine in which quarter degree grid squares (15’ × 15’) (QDGS) Blue Swallows have been recorded. This was used in conjunction with ArcView 3.2 to compile Figures 1–7.
The Blue Swallow African distribution range per quarter degree grid square based on distribution records 1850–2005. See the supplementary material for the references used to compile this map.
Blue Swallow distribution in the north-eastern DRC, Uganda, Kenya and northern Tanzania per quarter degree grid square based on distribution records 1850–2005. See Appendix 1 for the references used to compile this map.
Blue Swallow distribution in south-eastern DRC, southern Tanzania, Malawi, adjacent Mozambique and Zambia per quarter degree grid square based on distribution records 1850–2005. See the supplementary material for the references used to compile this map.
Blue Swallow distribution in Zimbabwe per quarter degree grid square based on distribution records 1850–2005. See the supplementary material for the references used to compile this map.
Blue Swallow distribution per quarter degree grid squares in South Africa and Swaziland based on all distribution records 1850–2005. See the supplementary material for the references used to compile this map.
Seven possible migratory routes used by Blue Swallows between their breeding and non-breeding grounds (also see Table 1).
The 13 quarter degree grid squares in which Blue Swallows were breeding in 2005 (Monadjem et al. Reference Monadjem, Boycott, Roques, Gama and Garcelon2006).
Migratory routes
The possible migratory routes used by Blue Swallows were largely determined by looking at the Blue Swallow breeding and non-breeding distribution ranges as they were in 1850 and classifying birds to be on passage to their non-breeding or breeding grounds if recorded outside of these areas in March–April and August–October, respectively. Information on the migratory patterns of other species was used (Berthold Reference Berthold1993) to compile a likely migratory scenario for the Blue Swallow.
Population decline
The South African and Swaziland distribution of the Blue Swallow based on QDGSs as it may have been in 1850 was compared with a map of the QDGSs known to contain breeding Blue Swallows in 2005. The reduction in QDGSs occupied by Blue Swallows over the past 155 years (1850–2005), and the current estimated population in South Africa and Swaziland, were used to estimate the 1850 population size for South Africa and Swaziland.
Literature was assembled on estimates of Blue Swallow populations from throughout its breeding and non-breeding range, and collated into a table. Estimates of Blue Swallow habitat surface area and the number of birds they contained were used to extrapolate the sub-population size for other population fragments. This was done for areas where they are known to occur and where there was some knowledge of suitable habitat, availability and extent. In this manner, an estimate of the entire Blue Swallow population was made per site occupied, and per country. This was used to assess the number of Blue Swallows that are protected (formally protected areas), partially protected (Ramsar Sites, Natural Heritage Sites, Forest Reserves, Important Bird Areas that are not formally protected areas), and unprotected, on both their breeding and non-breeding grounds.
Results
Distribution and area
Throughout their entire and exclusively African distribution range, Blue Swallows have been recorded between 26.11° to 36.95°E and 2.40°N to 30.60°S (Figure 1). On their non-breeding range, the Blue Swallows may have occupied as many as 23 QDGSs in the 1850s, with a combined surface area of approximately 16,100 km2 (Figures 1–2). On their breeding grounds, the Blue Swallows may have occupied as many as 138 QDGSs, in the 1850s, with a combined surface area of approximately 96,600 km2 (Figures 1 and 3–5). As each QDGS is about 700 km2 in size and Blue Swallows were probably not distributed throughout each QDGS, the above figures are probably an over-estimation of the 1850 range. However, based on this estimation, the original Blue Swallow breeding range was approximately 83% larger than their non-breeding range.
Migratory routes
Of the available records of Blue Swallows outside their usual breeding and non-breeding distribution ranges, it was judged that nine records were insufficiently or inaccurately dated for use in this analysis (Table 1). Based on the observation dates, six of the 11 accepted records of Blue Swallows outside their usual breeding and non-breeding ranges were of birds that may be on passage to their non-breeding grounds; the other five were probably birds on route to their breeding grounds, holding some clue as to the migratory routes used by Blue Swallows. Seven of these records are of single birds, while the others (two from Zimbabwe and one each from Malawi and Zambia) are of groups of birds. (Table 1, Figures 3 and 6).
Summary of the information available of all records of Blue Swallows considered out of range or on passage.
1 The locality reported for this skin should be accepted with reservation
2 Unconfirmed and not used further
3 The month was incorrectly documented as “13”, may be either the first or third month of the year?
4 11 records of birds possibly on migration and used in the analyses.
Figure 6 illustrates possible migratory routes that Blue Swallows may use, indicating three separate Blue Swallow sub-populations. There appears to be a disjuncture in the distribution of Blue Swallows between south-western and south-eastern Uganda. Some birds from the breeding grounds may migrate around the western side of Lake Victoria and others around the eastern. We therefore considered Lakes Victoria and Albert/Edward as barriers separating eastern and western migration routes and therefore sub-populations, and these sub-populations to fly east and west of Lakes Tanganyika and Malawi on migration. The first sub-population breeds in the south-eastern DRC and migrates (Route 1) to the north-eastern DRC (Lendu Plateau). The second sub-population breeds in South Africa, Zimbabwe, Mozambique, northern and south-western Malawi (Misuku Hills, Nyika National Park, North and South Viphya and Kirk Range), and south-western Tanzania (Umalali Mountains, Kitulo Plateau and Mount Rungwe) and migrates (Routes 2–5) to western Uganda and north-western Tanzania. The third breeds in south-eastern Malawi (Mount Mulanje) and southern Tanzania (Livingston Mountains and Udzungwa Mountains) and migrates (Routes 6 and 7) to Kenya and eastern Uganda. Except for the larger lakes, there appear to be no major mountain ranges or deserts that would act as barriers on the proposed routes.
Blue Swallows might migrate in groups. Just prior to migration, at the Blue Swallow Natural Heritage Site (South Africa), it has been observed towards the end of three Blue Swallow breeding seasons that Blue Swallows gather in groups (of up to 13) and then disappear, suggesting group migration (S. W. Evans pers. obs).
Population size and threat status
Since 1850 and before commercial forestry, Blue Swallows were documented occupying 49 QDGSs in South Africa and Swaziland (Allan et al. Reference Allan, Harrison, Navarro, Van Wilgen and Thompson1997). They were also recorded outside their normal breeding distribution range in a further seven QDGSs in South Africa. Based on historic (1850) and current (2005) distribution, the Blue Swallow’s range has declined by 74%, from 49 QDGSs to 13 (Figures 5 and 7). During this period in South Africa alone, there was a reduction from 41 QDGSs to 12, a 71% decline (Figures 5 and 6). During this period in Swaziland, the distribution range declined from eight QDGSs to one or possibly two, a 75–88% decline (Figures 5 and 7).
Assuming that the densities of Blue Swallows on their breeding grounds are similar to Malawi and Zimbabwe, the Blue Swallow population in the 1850s may have numbered 1,960–2,700 pairs (Table 2). Based on these estimations, the total current Blue Swallow population is calculated at approximately 1,006 pairs or 2,012 individual birds, a 36–56% decline from 1850 to 2005, compared with a decline of 74% for South Africa and Swaziland.
Estimate of the current total number of Blue SwallowsFootnote a.
a Barnes Reference Barnes1998, Bennun and Njoroge Reference Bennun and Njoroge1999, Reference Bennun, Njoroge, Fishpool and Evans2001, Childes in litt. 1999, 2000, Baker and Baker Reference Baker, Baker, Fishpool and Evans2001, Byaruhanga et al. Reference Byaruhanga, Kasoma and Pomeroy2001, Childes Reference Childes2001, Demey and Louette Reference Demey, Louette, Fishpool and Evans2001, Dowsett-Lemaire et al. Reference Dowsett-Lemaire, Dowsett, Dyer, Fishpool and Evans2001, Fishpool and Evans Reference Fishpool and Evans2001, Leonard Reference Leonard, Fishpool and I Evans2001, Parker Reference Parker, Fishpool and I Evans2001, Monadjem et al. Reference Monadjem, Boycott, Parker and Culverwell2003, Byaruhanga and Evans Reference Byaruhanga and Evans2004, Stjernstedt Reference Stjernstedt2004, EWT-Blue Swallow Working Group Monitoring records 2005.
As the Blue Swallow is a facultative co-operative breeder (Tree Reference Tree1989, Du Plessis et al. Reference Du Plessis, Siegfried and Armstrong1995, Wakelin Reference Wakelin2006), the 2,012 individual birds estimated for the current Blue Swallow population, based on active nest densities, may be an underestimation of the actual number of Blue Swallows, but not of active nests, as there may be more than two birds present at each active nest (Wakelin Reference Wakelin2006). Many Blue Swallows pairs are believed to breed co-operatively in KwaZulu-Natal with a mean of two females and one male per nest (Wakelin Reference Wakelin2006). At the study site in Mpumalanga containing nine pairs between 1995 and 1998, there was an equal sex ratio; cooperative breeding may have been taking place at only one nest (S. Evans unpubl. data).
Information on the longevity of swallows and martins that have northern and/or southern temperate and/or tropical zone breeding migrant populations that migrate into the range of their tropical breeding resident conspecifics (Banded Martin Riparia cincta, Lesser Striped-swallow Cecrops abyssinica, Red-breasted Swallow Cecrops semirufa and White-throated Swallow Hirundo albigularis) indicates that breeding longevity is approximately 1–7 years (McClure Reference McClure1974, Rydzewski Reference Rydzewski1978, Irwin Reference Irwin1981, Earlé Reference Earlé1987, Turner and Rose Reference Turner and Rose1989). The same was found for tropical breeding resident swallows and martins (Brown-throated Martin Riparia paludicola, Pacific Swallow Hirundo tahitica, and Ethiopian Swallow Hirundo aethiopica) (McClure Reference McClure1974, Rydzewski Reference Rydzewski1978, Irwin Reference Irwin1981, Earlé Reference Earlé1987, Turner and Rose Reference Turner and Rose1989). A conservative estimate of Blue Swallow longevity was therefore taken to be six years, and three generations was taken to span 18 years.
A decline in the Blue Swallow population in South Africa and Swaziland of 74% in 155 years represents a mean decline of 0.48% per year from 1850 to 2005. However, as commercial forestry commenced in Blue Swallow areas in the late 1880s (van der Zel Reference van der Zel1988) a 74% decline in 115 years (1890–2005) equates to a decline of 0.64% per year (Figures 5 and 7). A decline of 56% in the total Blue Swallow population over 155 years represents a mean decline of 0.36% per year from 1850 to 2005. Taking a cautious approach and therefore the worst-case scenario, the Blue Swallow qualifies as ‘Vulnerable’ under IUCN criterion C1 as the Blue Swallow population is estimated to have declined by as much as 12% over the next three generations. In addition, the Blue Swallow qualifies as Vulnerable under IUCN criterion C2a(i) due to a continuing decline projected in the number of mature individuals and no Blue Swallow sub-population estimated to contain more than 1,000 mature individuals IUCN/SSC (2000).
Protection status
Throughout their breeding range, approximately 60% of the Blue Swallow population is in strictly protected areas, a further 23% in partially protected areas (forest reserves, natural heritage sites and Ramsar sites), and an estimated 17% is unprotected (Table 3). The situation is unfortunately very different for the Blue Swallows when on their non-breeding range where only an estimated 28% of the population is in formally protected areas, an estimated 25% in partially protected areas, and an estimated 47% are unprotected (Table 3).
Current estimated Blue Swallow numbers and proportions that are protected throughout their breeding and non-breeding rangesFootnote a.
a Barnes Reference Barnes1998, Bennun and Njoroge Reference Bennun and Njoroge1999, Reference Bennun, Njoroge, Fishpool and Evans2001, Childes, pers. comm. 1999, 2000, Baker and Baker Reference Baker, Baker, Fishpool and Evans2001, Byaruhanga et al. Reference Byaruhanga, Kasoma and Pomeroy2001, Childes Reference Childes2001, Demey and Louette Reference Demey, Louette, Fishpool and Evans2001, Dowsett-Lemaire et al. Reference Dowsett-Lemaire, Dowsett, Dyer, Fishpool and Evans2001, Fishpool and Evans Reference Fishpool and Evans2001, Leonard Reference Leonard, Fishpool and I Evans2001, Parker Reference Parker, Fishpool and I Evans2001, Monadjem et al. Reference Monadjem, Boycott, Parker and Culverwell2003, Byaruhanga and Evans Reference Byaruhanga and Evans2004, Stjernstedt Reference Stjernstedt2004, EWT-Blue Swallow Working Group Monitoring records 2005.
Discussion
Population decline
The Blue Swallow population decline (Figures 5 and 7) in South Africa and Swaziland has primarily been due to a decrease in the grassland and wetland habitat quantity and quality available (Parker Reference Parker1994, Allan et al. Reference Allan, Harrison, Navarro, Van Wilgen and Thompson1997, Evans and Barnes Reference Evans, Barnes and Barnes2000, Monadjem et al. Reference Monadjem, Boycott, Roques, Gama and Garcelon2006). A reduction in habitat quantity and quality on the non-breeding grounds would also support a smaller population and result in there being fewer birds returning to South Africa and Swaziland to breed and therefore a decline in the overall Blue Swallow population (Evans et al. Reference Evans, Cohen, Sande, Monadjem, Hoffmann, Mattison, Newbery, Ndanganga and Friedmann2002).
The Blue Swallow population in Africa will continue to decline unless sufficient habitat is set aside to sustain viable populations. The International Blue Swallow Action Plan compiled in 2002 by representatives from nine of the ten Blue Swallow range states (Evans et al. Reference Evans, Cohen, Sande, Monadjem, Hoffmann, Mattison, Newbery, Ndanganga and Friedmann2002) needs to be reviewed, updated and implemented.
Important countries for conserving the Blue Swallow
The most important countries for conserving breeding populations of the Blue Swallow and its unique grassland and wetland habitats are Zimbabwe, Malawi and southern Tanzania, as combined they contain the largest numbers of breeding birds within the least fragmented habitat and contain 74% of the current estimated total breeding population (Tables 2 and 3). No dedicated survey of the Blue Swallow has been done in southern Tanzania and this is urgently needed.
When Blue Swallows are on their non-breeding range, Uganda and Kenya are the most important countries for conserving the birds and their wetland habitats, as these two countries currently contain an estimated 73% of the combined non-breeding Blue Swallow population.
Blue Swallow sub-populations
The Blue Swallow may be distributed within three separate sub-populations (Figure 6); it is therefore necessary to conserve the Blue Swallow on breeding and non-breeding grounds within all three separate sub-populations in order to conserve current genetic diversity. Blue Swallow stopover sites need to be identified along their migratory routes and programmes initiated to conserve them. Unfortunately, the impacts of changes along the Blue Swallow migratory routes are unknown and their effects on the population can not yet be determined. Studies are urgently needed to confirm the migratory routes used by Blue Swallows.
Protection status
Loss of habitat quality and quantity for Blue Swallows may be expected to be greatest in unprotected and partially protected areas, with very little to no habitat loss in formally protected areas. This would suggest that the current Blue Swallow population is at greatest risk on their non-breeding grounds where 47% of the population occurs in unprotected areas in Uganda and the Lendu Plateau in the north-eastern DRC.
Although this assessment was based on estimated numbers and distribution on their breeding and non-breeding range, it does provide a good indication that conservation of the Blue Swallow in Uganda and the Lendu Plateau in the DRC is important for securing its future survival.
Habitat on migratory routes
Six of the routes believed to be used by Blue Swallows are direct (Figure 6). It is believed that Blue Swallows follow Lake Tanganyika and the Albertine Rift as the leading lines aiding navigation between their breeding grounds in the south-eastern DRC and their non-breeding grounds on the Lendu Plateau. It was found, for instance, that migrants from North America migrating south either use the Rocky Mountains to their west or migrate south between the west coast and the Rocky Mountains to their east, as leading lines and aids to migration (Hynes Reference Hynes2007). A direct migratory route between breeding and non-breeding grounds need not be a straight line due to barriers such as mountain ranges and deserts that have to be crossed or avoided (Hynes Reference Hynes2007). The Purple Martin Progne subis breeds in open woodlands, farms and towns in North America, and migrates to similar habitat in South America during the non-breeding season (Turner and Rose Reference Turner and Rose1989). They must cross tropical forests on migration as there is no continuous open woodland linking North and South America (Turner and Rose Reference Turner and Rose1989). Aerial insectivores such as swallows and martins are able to forage on the wing and migrate by day, and when on migration, stop over near water, presumably with insects (Hynes Reference Hynes2007). This demonstrates that long-distance migrants such as the Purple Martin, and presumably the Blue Swallow, do not need habitats characteristic of their breeding or non-breeding habitat along the entire length of their migratory route. There are no major mountain ranges or deserts along the indicated routes (Figure 7). The 11 records of Blue Swallows outside their usual breeding and non-breeding distribution ranges were all within savannah habitat which is not typical of the grassland and wetland habitat characteristic of their breeding and non-breeding ranges (Figures 2–5 and 7).
Distribution
Sibley and Monroe (Reference Sibley and Monroe1990) erroneously included Lesotho in the southern African breeding distribution range of the Blue Swallow. Simmons and Barnard (Reference Simmons and Barnard2005) modelled Blue Swallow distribution from the distribution map and data in Harrison et al. (Reference Harrison, Allan, Underhill, Herremans, Tree, Parker and Brown1997). This model erroneously included south-western Mozambique, south-eastern Free State, the east of Eastern Cape Province and western Lesotho as part of the Blue Swallow distribution range. The distribution map in Stevenson and Fanshawe (Reference Stevenson and Fanshawe2002) includes Rwanda and northern Burundi in its non-breeding distribution range. In all three cases, there are no records supporting these distributions.
Conclusions
Numbers of Blue Swallows have declined by 56% over the last 155 years and only about 1,000 pairs are optimistically estimated to remain. With the low numbers to start with, and fewer still remaining, the survival of this species is in the balance. The Blue Swallow has lost large parts of its breeding range in southern Africa. Remnant Blue Swallow populations in southern Africa remain precarious. The remaining populations in Zimbabwe and Malawi now constitute the best survival option, but more could be done to formalise and upgrade the protection status of these areas. The non-breeding areas also need their protection status upgrading. The possible migratory routes need confirmation, as does the identification of any threats along these routes.
Supplementary Material
The supplementary materials referred to in this article can be found at journals.cambridge.org/bci
Acknowledgements
The Blue Swallow Working Group of the Endangered Wildlife Trust, North-West University, Paul Matiku, Achilles Byaruhanga, Warwick Tarboton, David Johnson, Peter Mundy, Bob Dowsett and Francois Dowsett-Lemaire are thanked for their contributions to this research.
